IoT Escalator Health Monitor with Vibration and Performance Dashboard

Escalators in malls and metro stations run 12–18 hours a day, and their failures are disruptive and occasionally dangerous — step-chain elongation, bearing wear and drive faults develop silently until a breakdown or an emergency stop. Maintenance today is mostly schedule-based: technicians inspect at fixed intervals whether the machine needs it or not. Vibration is the earliest tell — a degrading bearing or loosening chain changes the machine's vibration signature weeks before failure. This project builds the instrument that watches for it: a wireless ESP32 node with a magnet-mounted MEMS accelerometer at the drive-end bearing, plus motor current, step speed and gearbox temperature sensing, streaming to a dashboard that renders waveform and spectrum views, trends vibration RMS over 7 days, logs events, and suggests the right maintenance check in plain language. Strictly monitoring-only: nothing in this build can command the escalator.

1. The accelerometer is magnet-mounted at the drive-end bearing; current, speed and temperature sensors are fitted at the drive.
2. The ESP32 samples vibration at ~1.6 kHz, computing RMS, crest factor and dominant frequency per window, shown live on the OLED.
3. Feature sets plus current, speed and temperature are published over Wi-Fi to the backend with timestamps.
4. The dashboard renders live gauges, the waveform/spectrum view, the 7-day RMS trend and the event log.
5. Threshold and trend rules watch the data: RMS crossing the alert line or a rising elongation trend raises a fault event.
6. Each fault event carries a maintenance suggestion; the node itself only observes and never drives the escalator.

Tech stack:

• ESP32 (node controller + Wi-Fi)
• ADXL345-class MEMS accelerometer, magnetic mount
• Current · speed · temperature sensors
• 0.96" OLED on-node display
• Wi-Fi telemetry with timestamps
• Vibration feature extraction (RMS · crest factor · FFT)
• HTML · CSS · JS dashboard
• Threshold + trend alert rules engine

• [Vibration feature extraction] Tri-axial MEMS sensing with RMS, crest factor and dominant-frequency computed on the ESP32 at ~1.6 kHz sampling (design target).
• [Multi-channel monitoring] Motor current, step speed and gearbox temperature tracked alongside vibration for full context.
• [Waveform + spectrum views] Switchable time-waveform and frequency-spectrum plots of the latest vibration window on the dashboard.
• [7-day trend with alert line] Vibration RMS trended daily against a configurable alert threshold.
• [Event & fault log] Timestamped INFO/WARN events: comms gaps, heartbeat restores, threshold crossings.
• [Maintenance suggestions] Plain-language advisories from trend rules (e.g. chain-elongation trend → schedule tension check).
• [On-node OLED readout] Live vibration bar graph, RMS and status on the node's own display for field checks.

• Assembled sensor node (ESP32, accelerometer, OLED, sensor inputs, enclosure)
• Node firmware with vibration feature extraction and telemetry
• Web dashboard (gauges, waveform/spectrum, trend, event log, suggestions)
• Backend with data store and alert rules engine
• Installation and mounting notes
• Project report PDF (vibration theory, system design, methodology)
• PPT presentation and viva Q&A

• Vibration thresholds and sampling figures are design targets; fault rules are heuristic advisories, not certified diagnostics.
• The system must never be wired into escalator safety circuits or the controller — monitoring only, stated throughout the documentation.
• Installation needs facility permission and a safe mounting point; the guide covers placement, not site certification.
• Wi-Fi coverage is required at the node; a LoRa/cellular variant is future scope.
• Bearing-fault frequency matching is future scope — the base build reports features and trends, not named fault diagnoses.